Performanceproblems switching to faster server

[Ruecker, Rainer]

... just some additional info ...

I've run the same test with the config-files in a ram-disk
on both the p3 and the p4 - doesn't make a difference.
The trend is the same - P3 outperforms P4.

I made some more experiments with intels icc-6.0.1 and different CFLAGS for
gcc-3.1.1,
here are some resulte from the test-runs on the p4 box:
(Same setup as described in my previous mail...)

real    user   sys    compiler   flags
18,763  15,79  0,340  icc 6.0.1  -O2 -ipo -tpp7 -xW -fno-fnalias
17,687  16,79  0,114  gcc 3.1.1  -O -march=pentium4 -funroll-loops
-fno-strict-aliasing
17,080  16,349 0,127  gcc 3.1.1  -O3 -march=pentium4 -fno-strict-aliasing
-malign-double

So this is nagios doing the config-check. ( -v)
What happens on a multi-cpu-box if it starts to spawn the service-checks in
such
large scale installations as described by Marcus Hildenbrand might be
something
completely different. Any observations or comments?


Rainer Ruecker
-- 
STAG, Dircksenstrasse 42-44 10178 Berlin, Germany






> -----Original Message-----
> From: Ruecker, Rainer 
> Sent: Friday, October 25, 2002 11:55 PM
> To: 'nagios-devel at lists.sourceforge.net';
> 'nagios-users at lists.sourceforge.net'
> Subject: RE: [Nagios-users] Performanceproblems switching to faster
> server
> 
> 
> Hello,
> 
> I could not get my hands on our multiprocessor-boxes here 
> (yet) but did a
> few 
> tests on two single-cpu machines, just to see what's up with 
> the phenomen
> that
> has been described by Markus.
> 
> Machine 1: Intel(R) Pentium(R) 4 CPU 2.20GHz, cache size: 512 KB, 
>            1024 MB RAM - IDE Disk
>            Kernel Linux 2.4.18 sources from kernel.org
> 
> Machine 2: Intel(R) Pentium(R) III CPU 1400MHz, cache size: 512 KB,
>            512 MB RAM - SCSI Disk
>            Kernel Linux 2.4.7-10 RedHat
> 
> 
> Binaries from Nagios 1.0b5 have been compiled with the 
> following compilers
> on machine 1:
> 
> gcc version 2.95.3 20010315 (release)
> gcc version 2.96 20000731 (Red Hat Linux 7.1 2.96-98)
> gcc version 3.0.2 20010905 (Red Hat Linux 7.1 3.0.1-3)
> gcc version 3.1.1 (Thread model: posix)
> 
> .configure was called like this for each build:
> 
> ./configure  --prefix=/tmp/nag-[COMPILERVERSION] 
>              --with-command-grp=nobody 
>              --with-init-dir=/etc/init.d 
>              --with-lockfile=/var/lock/subsys/nagios 
>              --with-template-objects 
>              --with-file-perfdata 
>              --with-cgiurl=/nagios/cgi-bin 
>              --with-htmurl=/nagios 
>              --enable-embedded-perl
>              --with-perlcache
>              --with-template-extinfo
> 
> hosts.cfg    contains 1281 Hosts 
> services.cfg contains 6379 Services
> (There are 1272 Hosts that have all the same 5 service-checks ...)
> 
> Here are the average runtimes, taken from 30 runs of "time 
> bin/nagios -v
> etc/nagios.cfg"
> for each build 
> 
>                 gcc295             gcc296             gcc301
> gcc311
>             real user sys      real user sys      real user 
> sys      real
> user sys          
>  
> Maschine 1 17,87 15,44 0,332  19,74 16,59 0,360  18.31 15,77 
> 0,344  18,38
> 15,90 0,338
> 
> Maschine 2  9,81  9,77 0,039   9,96  9,90 0,036      libc6-error
> libc6-error
> 
> No Typo - I suppose it could be the disk subsystem and will 
> investigate
> further.
> 
> If someone needs the complete results from the 240 test-runs 
> or needs more 
> details about the setup, just drop me a mail.
> 
> Cheers
> 
> Rainer Ruecker
> -- 
> STAG, Dircksenstrasse 42-44 10178 Berlin, Germany
> 
> 
> > -----Original Message-----
> > From: Marcus Hildenbrand [mailto:marcus.hildenbrand at sap.com]
> > Sent: Tuesday, October 15, 2002 12:43 PM
> > To: nagios-devel at lists.sourceforge.net;
> > nagios-users at lists.sourceforge.net
> > Subject: [Nagios-users] Performanceproblems switching to 
> faster server
> > 
> > 
> > Our current produktive Nagios installation runs on a 4x700 
> MHz server 
> > with Suse Linux. We are doing an active monitoring of > 
> 1.100 servers 
> > with > 5.500 services. Because of the known problems with 
> > slow CGI's on 
> > large configurations we decided to test our Nagios 
> installation on a 
> > 2x2.4 GHz Suse Linux server hoping to get the CGI's faster. 
> > But we saw 
> > the opposite. For example the command "time 
> > <path_to_nagios>/bin/nagios 
> > -v <main_config_file>" gives the following response:
> > 
> > 
> > 4x700 MHz server:
> > ..
> > ..
> > Things look okay - No serious problems were detected during the 
> > pre-flight check
> > 
> > real    0m1.866s
> > user    0m1.820s
> > sys     0m0.050s
> > 
> > 
> > 
> > 2x2.4 GHz server:
> > ..
> > ..
> > Things look okay - No serious problems were detected during the 
> > pre-flight check
> > 
> > real    0m7.428s
> > user    0m7.410s
> > sys     0m0.020s
> > 
> > Has anyone an idea why this is more than 3 times slower on 
> the faster 
> > box? This problem also exists for the CGI's.
> > 
> > Thanks Marcus
> > 
> > 
> > 
> > -------------------------------------------------------
> > This sf.net email is sponsored by:ThinkGeek
> > Welcome to geek heaven.
> > http://thinkgeek.com/sf
> > _______________________________________________
> > Nagios-users mailing list
> > Nagios-users at lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/nagios-users
> > 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by: Influence the future 
> of Java(TM) technology. Join the Java Community 
> Process(SM) (JCP(SM)) program now. 
> http://ads.sourceforge.net/cgi-bin/redirect.pl?sunm0004en
> _______________________________________________
> Nagios-users mailing list
> Nagios-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/nagios-users
> 


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en